[lammps-users] Problems with wurtzite modification of ZnS

Dear LAMMPS Users,

I contact you since one of my students has come across some problems simulating ZnS. After creating a wurtzite lattice and implementing a Buckingham potential from literature we saw that the stress distribution (pxx, pyy, pzz) was not isotropic and the system pressure not 0 (or at least fluctuating around 0). We’ve then decided to relax the structure by creating a supercell in conjunction with a berendsen barostat and the “aniso” option activated. The supercell was optimized using several series alternating each between minimization steps and short dynamic runs (NVE). Finally the supercell was scaled to +1% of its initial size using the “deform” command and then compressed to -1% in order to calculate the bulk modulus.

While the minimization + berendsen approach delivered a structure at around 0 bar in all directions (and minor fluctuations), the scaling of the system lead to a rather strange behaviour (se attached plot). Usually I would expect a U-shaped curve with a minimum at the equilibrium distance.

I wanted to ask if anyone has encountered such behaviour before or knows how to interpret it? Also can someone give us a hint how to create a stable wurtzite structure without the usage of minimization or the steps described above? I haven’t had such problems with previous materials.

I have attached the script in a simplified form at the end of this mail. Perhaps you can have a look at it.

Thank you in advance for your help!

Best regards,

Patrick

LAMMPS Script (shortened):

ZnsWurt_CellOpt_BulkMod.jpg